鲁棒的混合域音频信息隐藏算法

提出了一种基于离散小波变换(DWT)和离散余弦变换(DCT)的音频信息隐藏算法。对载体音频整体进行小波分解,将其划分成若干频带,然后根据人耳听觉的频率掩蔽效应,选择对人耳听觉不敏感的频带所对应的小波系数,将小波系数分段进行离散余弦变换,将水印嵌入到DCT直流系数上。实验表明,嵌入水印后的音频文件不仅具有良好的不可感知性,而且对诸如噪声、低通滤波、重采样、回声和Mp3压缩等的攻击具有很强的鲁棒性,并能抵抗一定程度的样点裁剪攻击,算法的鲁棒性和不可感知性达到了很好的平衡。

Robust Audio Information Hiding Algorithm Based on DWT and DCT

A new audio information hiding algorithm based on the DWT and DCT domain is proposed.First the entire host audio is decomposed by wavelet to several frequency bands,then select a frequency band which is insensitivity to the human ear hearing and its corresponding wavelet coefficients to hide information.Those coefficients are separated to some segments,and at the same time convert those segments to DCT domain.The watermark can be embedded to the DCT direct current coefficients.Experimental results show that the watermarked audio has good imperceptibility and is robust against different kinds of attacks,such as samples cropping,noise adding,low-pass filtering,resampling,echo,Mp3 compression.The robustness and the imperceptibility about watermarked audio reach a good balance.